This resubmission K23 proposal outlines a five-year development plan for Sophia Koo to achieve independence as a patient-oriented clinician-scientist with particular expertise in the development and assessment of novel diagnostic modalities for invasive mold infections (IMI). The strategy outlined in this proposal will enable Dr. Koo to build upon her clinical background in caring for patients at risk for IMI and her clinicl research defining the limitations of currently available diagnostic modalities for these infections by taking advantage of the unique clinical and research strengths of Brigham and Women's Hospital and the Dana-Farber Cancer Institute, biomedical engineering and bioinformatics capabilities of the Charles Stark Draper Laboratory, and epidemiology, biostatistics, and analytical chemistry training opportunities at the Harvard School of Public Health, Northeastern University, and Massachusetts Institute of Technology. She will be co-mentored by Dr. Lindsey Baden, Associate Professor of Medicine at Harvard Medical School, and Dr. Earl Francis Cook, Professor of Epidemiology at the Harvard School of Public Health. Dr. Baden is an expert in fungal diagnostics, and Dr. Cook is an expert in study design and statistical learning/data mining, including analysis of highly dimensional biomarker data. Dr. Koo will also work closely with Drs. Daniel Kuritzkes, Raphael Dolin, Francisco Marty, John Perfect, and John Wingard, who will provide career guidance and scientific advice on the execution of the proposed research plan, and her collaborators Drs. Paul Vouros, Terence Risby, and James Comolli, who will contribute their expertise in analytical chemistry and breath analysis to her research plan. The research program will employ analytical chemistry and statistical supervised learning approaches to define volatile organic compound (VOC) metabolite patterns of emerging invasive mold species in vitro and in patient breath to lay the groundwork for a noninvasive, point-of-care breath test for patients with emerging IMI. Specifically, this program proposes (1) a detailed definition of the VOC metabolome of the Mucorales, Fusarium, and Scedosporium species that most commonly cause human IMI, and (2) the prospective collection of VOC metabolites in the breath of patients with emerging IMI, patients with clinical and radiographic findings suggestive of an emerging IMI, and patients with respiratory tract colonization by these molds, with derivation and validation of VOC feature support vector machine classification algorithms to discriminate patients with emerging IMI from patients with other fungal infections or non-fungal pulmonary processes. Current diagnostic modalities for these infections are inadequate and often unacceptably morbid, and development of this novel, noninvasive diagnostic approach should allow earlier diagnosis of these infections than currently possible, prospective surveillance in the growing population of patients at risk for emerging IMI, more rational targeting of antifungal therapy, and improved clinical outcomes. Project Narrative Invasive mold infections are highly morbid and often fatal in the growing population of patients with compromised immune systems. The proposed research seeks to overcome the significant limitations of currently available diagnostic methods for these infections by developing a novel, noninvasive test that identifies unique volatile compounds in the breath of patients with these infections. Development of this sensitive fungal breath test will allow better surveillance and targeting of antifungal therapy in patients at risk for these infections and ultimately, improved patient outcomes.